bioRxiv preprint doi: https://doi.org/10.1101/254664; this version posted January 27, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Title Experimental analysis of exome-scale mutational signature of glycidamide, the reactive metabolite of acrylamide Authors Maria Zhivagui1, Maude Ardin1, Alvin W. T. Ng2,3,4, Mona I. Churchwell5, Manuraj Pandey1, Stephanie Villar1, Vincent Cahais6, Alexis Robitaille7, Liacine Bouaoun8, Adriana Heguy9, Kathryn Guyton10, Martha R. Stampfer11, James McKay12, Monica Hollstein1,13,14, Magali Olivier1, Steven G. Rozen2,3, Frederick A. Beland5, Michael Korenjak1 and Jiri Zavadil1 Affiliations 1 Molecular Mechanisms and Biomarkers Group, International Agency for Research on Cancer, Lyon 69008, France 2 Centre for Computational Biology, Duke-NUS Medical School, Singapore 169857, Singapore 3 Program in Cancer and Stem Cell Biology, Duke-NUS Medical School, 169857, Singapore 4 NUS Graduate School for Integrative Sciences and Engineering, 117456, Singapore 5 Division of Biochemical Toxicology, National Center for Toxicological Research, Jefferson, AR 72079, USA 6 Epigenetics Group, International Agency for Research on Cancer, Lyon 69008, France 7 Infections and Cancer Biology Group, International Agency for Research on Cancer, Lyon 69008, France 8 Environment and Radiation Section, International Agency for Research on Cancer, Lyon 69008, France 9 Department of Pathology and Genome Technology Center, New York University, Langone Medical Center, New York, NY 10016, USA 10 IARC Monographs Section, International Agency for Research on Cancer, Lyon 69008, France 11 Biological Systems and Engineering Division, Lawrence Berkeley National Laboratory, Berkeley, CA, 94720, USA 12 Genetic Cancer Susceptibility Group, International Agency for Research on Cancer, Lyon 69008, France 13 Deutsches Krebsforschungszentrum, 69120 Heidelberg, Germany 14 Faculty of Medicine and Health, University of Leeds, LIGHT Laboratories, Leeds LS2 9JT, United Kingdom Keywords: Acrylamide, glycidamide, DNA adducts, massively parallel sequencing, mutational signatures Correspondence: [email protected] and/or [email protected] 1 bioRxiv preprint doi: https://doi.org/10.1101/254664; this version posted January 27, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Abstract Acrylamide, a probable human carcinogen, is ubiquitously present in the human environment, with sources including heated starchy foods, coffee and cigarette smoke. Humans are also exposed to acrylamide occupationally. Acrylamide is genotoxic, inducing gene mutations and chromosomal aberrations in various experimental settings. Covalent haemoglobin adducts were reported in acrylamide-exposed humans and DNA adducts in experimental systems. The carcinogenicity of acrylamide has been attributed to the effects of glycidamide, its reactive and mutagenic metabolite capable of inducing rodent tumors at various anatomical sites. In order to characterize the pre-mutagenic DNA lesions and global mutation spectra induced by acrylamide and glycidamide, we combined DNA-adduct and whole-exome sequencing analyses in an established exposure-clonal immortalization system based on mouse embryonic fibroblasts. Sequencing and computational analysis revealed a unique mutational signature of glycidamide, characterized by predominant T:A>A:T transversions, followed by T:A>C:G and C:G>A:T mutations exhibiting specific trinucleotide contexts and significant transcription strand bias. Computational interrogation of human cancer genome sequencing data indicated that a combination of the glycidamide signature and an experimental benzo[a]pyrene signature are nearly equivalent to the COSMIC tobacco-smoking related signature 4 in lung adenocarcinomas and squamous cell carcinomas. We found a more variable relationship between the glycidamide- and benzo[a]pyrene-signatures and COSMIC signature 4 in liver cancer, indicating more complex exposures in the liver. Our study demonstrates that the controlled experimental characterization of specific genetic damage associated with glycidamide exposure facilitates identifying corresponding patterns in cancer genome data, thereby underscoring how mutation signature laboratory experimentation contributes to the elucidation of cancer causation. A 40-word summary Innovative experimental approaches identify a novel mutational signature of glycidamide, a metabolite of the probable human carcinogen acrylamide. The results may elucidate the cancer risks associated with exposure to acrylamide, commonly found in tobacco smoke, thermally processed foods and beverages. 2 bioRxiv preprint doi: https://doi.org/10.1101/254664; this version posted January 27, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. Introduction Cancer can be caused by chemicals, complex mixtures, occupational exposures, physical agents, and biological agents, as well as lifestyle factors. Many human carcinogens show a number of characteristics that are shared among carcinogenic agents (1). Different human carcinogens may exhibit a spectrum of these key characteristics, and operate through separate mechanisms to generate patterns of genetic alterations. Recognizable patterns of genetic alterations or mutational signatures characterize carcinogens that are genotoxic. Recent work shows that these DNA sequence changes can be expressed in simple mathematical terms that enable mutational signatures to be extracted from thousands of cancer genome sequencing data sets (2). Several of the over 30 identified mutational signatures have been attributed to specific external exposures or endogenous factors through epidemiological and experimental studies (2). However, about 40% of the current signatures remain of unknown origin, and additional, thus far unrecognized, signatures are likely to be defined in rapidly accumulating cancer genome data. Well-controlled experimental exposure systems can thus help identify the underlying causes of known orphan mutational signatures as well as define new patterns generated by candidate carcinogens (reviewed in (3,4)). Various diet-related exposures contribute to the human cancer burden. Examples include contaminants in food or alternative medicines, such as aflatoxin B1 (AFB1) or aristolochic acid (AA). The mutagenicity of these compounds is well-documented; AFB1 induces predominantly C:G>A:T base substitutions and AA causes T:A>A:T transversions. The characteristic mutations coupled with information on the preferred sequence contexts in which they are likely to arise allowed unequivocal association of exposure to AFB1 or AA with specific subtypes of hepatobiliary or urological cancers, respectively (5-13). Among dietary compounds with carcinogenic potential, acrylamide is of special interest due to extensive human exposure. Important sources of exposure to acrylamide include tobacco smoke (14), coffee (15), and a broad spectrum of occupational settings (16). Dietary sources of acrylamide comprise carbohydrate-rich food products that have been subject to heating at high temperatures. This is due to Maillard reactions, which involve reducing sugars and the amino acid asparagine, present in potatoes and cereals (17). There is sufficient evidence that acrylamide is carcinogenic in experimental animals (18,19) and it has been classified as a probable carcinogen (Group 2A) by the International Agency for Research on Cancer in 1994 (16). The association of dietary acrylamide exposure with renal, endometrial and ovarian cancers has been explored in recent epidemiological studies (20,21). However, accurate acrylamide exposure assessment in epidemiological studies based on questionnaires has been difficult, and more direct measures of molecular markers, such as hemoglobin adduct levels, may not yield conclusive findings on past exposures (22- 3 bioRxiv preprint doi: https://doi.org/10.1101/254664; this version posted January 27, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 27). An improved understanding of its mechanism of action using well-controlled experimental systems is critical for understanding the potential carcinogenic risk associated with exposure. Acrylamide undergoes oxidation by cytochrome P450, producing the reactive metabolite glycidamide that is highly efficient in DNA binding due to its electrophilic epoxide structure (28-30). The Hras mutation load in neoplasms of mice exposed to acrylamide or glycidamide was found to be considerably higher in mice treated with glycidamide (31). This finding is corroborated by a considerably higher mutation frequency in the cII reporter gene of Big Blue mouse embryonic fibroblasts treated with glycidamide in comparison to acrylamide (32,33). Mutation analysis in different experimental in vivo and in vitro models using reporter genes showed an increased association